JP4229325B2 - Peak detection image processing method, program, and apparatus - Google Patents

Description

  The present invention relates to a grayscale image data group comprising grayscale data representing gray levels of each section obtained by dividing a grayscale image corresponding to a distribution in a space of an arbitrary physical quantity into a plurality of small sections in association with position data representing the position of each section. In particular, the present invention relates to a peak detection image processing method for detecting a peak value of grayscale data, its position and the number of peaks.

  The present invention is based on image processing / analysis from (1) the quantity of powder, particles, or an image obtained by image measurement of powder produced as an industrial product or a powder in which many fine particles are collected, particulate matter, etc. It can be used when (2) powder, particle size, particle diameter, and (3) particle position are detected from the luminance of the particles in the image. Or, in the field of environmental science research such as the air purification industry, the present invention is used to determine the number, particle size, and position in the image from a spotted gray image obtained by observing fine particles floating in the atmosphere or liquid. can do. Furthermore, it can also be used for inspection of irregularities on a flat plate such as a semiconductor silicon wafer and detection of dust such as fine dust adhering to them.

  In the conventional technique, in order to accurately detect or obtain the peak value and the number of spots in a gray-scale image in which a large number of spots as the subject of the present invention are collected, generally, there are a number of special and complicated processes. It uses a lot of filtering. For example, first, an image is smoothed by processing the image with a low-pass characteristic filter such as a moving average method or a median filter to create an envelope image. Next, a difference from the original image is taken, and a process of extracting only spots or dot-like portions is performed. Then, after binarizing the difference image by threshold processing on the difference image, the thinning process is repeated a plurality of times to cut down to one pixel. Finally, it is necessary to obtain a number of pixels having a peak and a peak value by counting the pixels that have become one pixel. At this time, when the peak value is acquired for each spot, a process for leaving the maximum value (peak value) of the spot area in the difference image is required.

Examples of the prior art related to the present invention include the following.
JP 2002-262093 A JP 2003-259129 A K. R. Cast Luma, translated by Tsuneo Saito, "Digital Image Processing", Science and Technology Publishing, published in March 2002 A. Rosenfeld, translated by Shin Nagaga, "Digital Image Processing", Modern Science, published in December 1978 Satoshi Kawada / Shigeo Minami, "Image data processing for scientific measurement", CQ Publisher, published April 30, 1994

  When obtaining the peak value and the number of spots in an image using the conventional image processing method, a plurality of image processing methods such as filtering, binarization, thinning, etc. must be repeatedly used as described above. . For this reason, there arises a problem that the procedure and process of image processing become complicated and a huge amount of image processing time is required. Furthermore, the raw peak value is not stored and maintained by a number of image processing, and may change.

  The present invention provides a plurality of image processing methods in image processing for obtaining peak values and positions of a large number of spots in a grayscale image corresponding to a distribution in a space of an arbitrary physical quantity, and the number of peak values if desired. It is an object of the present invention to solve the two major problems of the application procedure and the complexity of the processing process and the enormous image processing time required for the processing so that the processing can be performed easily and in a short time.

  In order to solve the above-described problem, the peak detection image processing method according to the present invention is configured so that the grayscale data representing the grayscale of each section obtained by dividing the grayscale image corresponding to the distribution in the space of an arbitrary physical quantity into a plurality of small sections is stored in each section. A step of individually comparing the grayscale data value of one section in the grayscale image data group included in association with the position data representing the position with the respective grayscale data values of all surrounding sections surrounding the one section; When the data value of the one partition is greater than or equal to all the data values of the surrounding partition, or when the data value of the one partition is less than or equal to all the data values of the surrounding partition, or both A peak determining step of determining a data value of the one section as a peak value and determining a position of the one section as a peak position. To.

It is preferable that the comparison calculation step and the peak determination step are executed for all sections.
In one aspect of the present invention, the grayscale image is a two-dimensional rectangular image, and the section is formed by dividing the rectangular image into at least three parts in the vertical and horizontal directions, and the comparison operation The step of comparing the data value of the one section with the data values of the eight sections surrounding the one section when the one section is a section that is not in contact with the outer periphery of the rectangular image; When one section is a section that is in contact with the outer periphery of only one of the vertical and horizontal directions of the rectangular image, the data value of the one section is used as data of five sections surrounding the one section. When the one section is a section at the corner of the rectangular image, the data value of the one section is preferably compared with the data values of the three sections surrounding the one section. .

  The peak detection image processing method of the present invention further selects one of the plurality of sections using an arbitrary selection method when a plurality of sections having data values determined as peak values are adjacent to or adjacent to each other. Preferably, the method includes the step of:

The peak detection image processing method of the present invention preferably further comprises a step of counting the number of peak values determined by the peak determination step.
In another aspect of the present invention, the grayscale image is optically acquired by an optical image detector such as an image pickup tube or a CCD, and the grayscale data and position data of the grayscale image data group are the acquired grayscale. It is preferably generated as a digital value from the image.

Preferably, the peak detection image processing method of the present invention further includes a step of recording the determined peak value and peak position.
The above problem is solved by a peak detection image processing program that causes a computer to execute the above-described various peak detection image processing methods.

  In order to solve the above-described problem, the peak detection image processing apparatus according to the present invention converts the grayscale data representing the grayscale of each section obtained by dividing the grayscale image corresponding to the distribution in the space of an arbitrary physical quantity into a plurality of small sections. Means for individually comparing and calculating the grayscale data values of one section in the grayscale image data group included in association with the position data representing the position with the respective grayscale data values of all surrounding sections surrounding the one section; When the data value of the one partition is greater than or equal to all the data values of the surrounding partition, or when the data value of the one partition is less than or equal to all the data values of the surrounding partition, or both In the case of the above, it is characterized by comprising peak determining means for determining the data value of the one section as a peak value and determining the position of the one section as a peak position.

  By providing the above-described configuration, the present invention can easily and quickly detect the peak value of grayscale data and its position in a grayscale image corresponding to the distribution of an arbitrary physical quantity in a space.

  Further, the present invention has the above-described configuration, so that various image processing and complicated processes that have been indispensable for detection and analysis of a large number of peak values in the conventional image are unnecessary, and a large amount of processing is required. Calculation time can be greatly reduced. Further, it becomes possible to detect and analyze the peak value in the image in real time, which was impossible in the past.

For example, in the case of an image having many spots in a grayscale image, the peak value of each spot in the image, the number of pixels having a peak, and the peak position on the screen can be obtained simultaneously.
In one aspect of the present invention, when a pixel having the same peak value is adjacent and forms a continuous area on the screen, an arbitrary method or criterion for selecting one of the pixels is added. Thus, the peak value can be narrowed down to one, and the number of spots having the peak value can be obtained accurately.

Preferred embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a block diagram schematically showing the configuration of a preferred embodiment of the peak detection image processing apparatus of the present invention. The peak detection image processing apparatus is acquired by an image input unit 10 and an image input unit 10 that receive an image optically acquired by an optical image detector such as an imaging tube or a CCD (charge coupled device) camera in a digital format, for example. A digital image and a peak detection image processing program for detecting a light and dark peak in the digital image, and an external storage device 12 including a work area being processed, and called from the external storage device 12 for image processing. A main memory 14 for storing the peak detection image processing program and image data to be processed or data generated during the processing, and a central processing unit for executing the peak detection processing according to the peak detection image processing program of the main memory 14 ( CPU) 16, for inputting control information necessary to execute the peak detection process Input device 18, and intermediate results of the peak detection processing such Bodo, final result, a display 20 for displaying the input control information. These components are interconnected via a normal bus.

Next, preferred embodiments of the peak detection image processing method and program of the present invention will be described below with reference to the drawings.
The grayscale image that is the processing target of the present invention, for example, irradiates fine particles floating in an arbitrary indoor space with laser light as described later with reference to FIG. This is a grayscale image obtained by photographing with several 324,064 CCD cameras. The shading of each pixel in the shading image in FIG. 4 corresponds to the intensity of scattered light, and the digital value representing the shading of each pixel represents the intensity of scattered light. The grayscale image to be processed according to the present invention is not limited to an image obtained by photographing with such a CCD camera or the like, and any arbitrary means distributed in two dimensions or three dimensions using any means. Any object may be used as long as it is obtained by imaging a target and the resulting image is expressed in shades. The grayscale image may be in direct digital form as in the case of a CCD, or in analog form as in the case of a normal camera. In the latter case, any means for converting the analog data of the grayscale image into digital data may be provided externally or in the peak detection image processing apparatus of the present invention, and the analog value may be converted into a digital value by the means. Further, the image input unit 10 may receive image data sent from another arbitrary system directly or via a communication medium such as a communication line.

  The peak detection image processing method according to the present invention is applied to sample matrix data of n rows and m columns as shown in FIG. 2 acquired as spot-like image data in which a large number of spot or dot-like images are collected in the image. Note that n and / or m is an integer of 2 or more, and preferably 3 or more. FIG. 3 shows a peak detection image processing flow. In the procedure of the peak detection image processing, as shown in the peak detection image processing flow of FIG. 3, at the same time as the processing is started, image matrix data to be processed is selected in step S1. Next, in step S2, a pixel data value x (i, j) in any one i row and j column that has not yet been selected for peak detection image processing, that is, has not yet been compared, is selected. Next, in step S3, the selected pixel data value x (i, j) is converted into eight neighboring pixel data values x (i-1, j-1), x (i-1, j), x ( i-1, j + 1), x (i, j-1), x (i, j + 1), x (i + 1, j-1), x (i + 1, j), x (i + 1, j + 1) and when the pixel data value x (i, j) satisfies the following relational expression (1) with the surrounding pixel data value, that is, compared with each surrounding pixel data value When this pixel data value x (i, j) is greater than all the surrounding pixel data or equal to the largest of the surrounding pixel data, the pixel data value x (i, j) is detected as a peak value In step S4, the pixel data value x (i, j) having a peak value is recorded and stored in the external storage device 12.

  On the other hand, if the pixel data value x (i, j) is smaller than any of the surrounding pixel data values in step S3, the pixel data value is not recorded or stored as pixel data having a peak value. The peak detection image processing then proceeds to step S5, where it is determined whether all image data (n × m) of the image matrix data of n rows and m columns has been processed. If there is unprocessed image data, the process returns to step S2 and the above-described processing is repeated. If it is determined in step S5 that all image data has been processed, the peak detection image processing ends.

  The peak detection image processing method of the above embodiment is simply described. For all the pixel data of the image matrix data, the data value of any one pixel is set around the pixel or adjacent pixel data values one by one. This is a processing method of finding a pixel having a data value higher than the data value of surrounding pixels and a peak value (local maximum value) by performing comparison operation on the data of all pixels.

  By applying the above processing to all the pixel data of the matrix data, it becomes possible to determine the peak value of the spot or point pixel in the image, and at the same time, the number and the position of the pixel are also determined.

  The pixel data x (1,1), x (n, 1), x (1, m), and x (n, m) located at the corners of the image matrix data of n rows and m columns shown in FIG. In some cases, each image data is compared with three pixel data surrounding those pixels. For example, in the case of x (1,1), comparison processing is performed with x (2,1) and x (1,2) x (2,2), respectively. Further, one row 2 rows 1 column x (2,1) to n-1 rows 1 column x (n-1,1), m rows 2 rows m columns x (2, m) to n-1 rows m columns x (n-1, m), 1 row 1 row 2 columns x (1,2) to 1 row m-1 column x (1, m-1), and n rows In the case of pixel data from n rows and 2 columns x (n, 2) to n rows and m-1 columns x (n, m-1), each image data is compared with the five pixel data surrounding them. To do.

  As a result of image processing, x (i, j), x (i, j + 1), x (i + 1, j), and x (i + 1, j + 1) have the same value in FIG. If all four adjacent pixels have the same spot or speckle peak value, it is recorded and saved as a peak value by any method. By doing so, the number of peaks can be obtained accurately. For example, in such a case, if a method for selecting a pixel having a large number of rows and columns is added, x (i, j), x (i, j + 1), x (i + 1, j) , X (i + 1, j + 1) of x (i + 1, j + 1) is recorded and stored in the external storage device 12 as a pixel having a peak value, and the number of pixels having a peak is accurately obtained. It becomes possible to do.

Next, a measurement example or experimental data when the peak detection image processing flow shown in FIG. 3 is applied to an image obtained by irradiating fine particles floating in an arbitrary indoor space with laser light and capturing the scattered light will be described. . FIG. 4 shows an image obtained by irradiating fine particles floating in an arbitrary space in a room with laser light and capturing scattered light from the measured fine particles with a CCD camera having 324,064 pixels. A large number of spots or dot spots in the image are scattered light from each fine particle in the air. This image data is matrix data of 494 rows and 656 columns. A part thereof is shown in FIG. The intensity of the scattered light is represented by black and white shading on the screen, where the darkest part is the weakest scattered light intensity and the whitest part is the strongest scattered light intensity. FIG. 5 shows the density of each pixel of the image shown in FIG. 4 as a relative value, assuming that the state of only the dark current of the CCD having the minimum intensity is 0 and the maximum intensity is 2 ¹² = 4096.

  FIG. 6 shows the matrix data of FIG. 5 in a three-dimensional graph. One of the many mountain-shaped or steeple-shaped ones shown in FIG. 6 shows scattered light from one fine particle, which represents one spot in FIG. The number of peaks represents the number of fine particles, and the height and peak value indicate the size of the fine particles.

  The matrix data in FIG. 7 is obtained by applying the peak detection image processing method of the present invention to the matrix data in FIG. 5 which is the image data in FIG. In the matrix data of FIG. 7, for convenience, the matrix data of pixels other than the peak value is displayed as 0, and the peak value is described only in the matrix data corresponding to the pixel having the peak value. Thereby, the number of pixels having a peak value, the peak value, and the position of the pixel on the screen can be obtained.

  If desired, the matrix data after this image processing is collectively shown in a histogram (frequency distribution graph) in FIG. 8 and the peak value included in each range of peak values is shown in FIG. A table can be easily obtained.

  The peak detection image processing apparatus of the present invention is not limited to the configuration shown in FIG. 1, and the above-described peak detection image is used by using an arbitrary type of computer such as a normal personal computer, a microprocessor, a workstation, or a general-purpose computer. A processing program can be executed.

  The present invention is not limited to the case where the peak value is a maximal value as in the above-described embodiment. In the comparison operation, the data value of the surrounding pixels is equal to the smallest data value or smaller than all the data values. In this case, by repeating the operation of recording / storing the pixel as the data value x (i, j) having the minimum value, the luminance distribution depressed on the minus side in the image is detected in the same manner as the detection of the pixel having the peak value. It is also possible to acquire the number of minimum values (negative peak values) of the spots, the peak value, and the position on the screen.

  In the above embodiment, the case where the outer shape is rectangular as shown in FIG. 4 as a gray image has been described. However, in the present invention, the shape of the gray image is not limited to a rectangle, and any two such as a circle may be used. It may be a dimensional shape, and it is only necessary to specify the position of a section (corresponding to the pixel in the above embodiment) obtained by dividing the gray image into small sections. Moreover, a division is not limited to the rectangle of said embodiment, Arbitrary shapes may be sufficient if the magnitude | size of a division is uniform.

  Furthermore, in the above embodiment, the grayscale image is a two-dimensional image as shown in FIG. 4, but the present invention can also be applied to a case where the grayscale image has a three-dimensional configuration. In that case, as a section, for example, a small three-dimensional gray image is divided into small cubes, and the peak of the data value corresponding to the light and shade of the cube is set in a three-dimensional manner in the same manner as in the above embodiment. What is necessary is just to detect by comparing with the data value of the adjacent cube which encloses, and to specify the position of the cube.

FIG. 1 is a block diagram schematically showing the configuration of a preferred embodiment of the peak detection image processing apparatus of the present invention. FIG. 2 shows sample matrix data of n rows and m columns acquired as spot-like image data in which many spots or dot-like images are collected in the image. FIG. 3 shows a peak detection image processing flow according to a preferred embodiment of the present invention. FIG. 4 shows an image obtained by irradiating fine particles floating in an arbitrary space in a room with laser light and capturing scattered light from the measured fine particles with a CCD camera having 324,064 pixels. FIG. 5 shows sample matrix data representing an arbitrary part of the image shown in FIG. FIG. 6 shows a three-dimensional distribution map of an arbitrary portion of the scattered light image of FIG. 4 or FIG. FIG. 7 shows matrix data obtained by applying the peak detection image processing method of the present invention to the matrix data of FIG. 5 which is the image data of FIG. FIG. 8 shows a histogram representing the number and frequency of the peak values after the image processing shown in FIG. FIG. 9 is a chart showing the number of peaks included in the range of the peak value range of the data shown in FIG.

Explanation of symbols

10 Image Input Unit 12 External Storage Device 14 Main Memory 16 Central Processing Unit (CPU)
18 Input device 20 Display

Claims (10)

1 in a grayscale image data group including grayscale data representing gray levels of each section obtained by dividing a grayscale image corresponding to a distribution in a space of an arbitrary physical quantity into a plurality of small sections in association with position data representing the position of each section. Individually comparing gray data values of one section with the gray data values of all surrounding sections surrounding the one section;
Said one if the data value of the partition is greater than or equal to all of the data values of the surrounding compartments, or the one compartment of the data values of all the data values smaller than or equal if the surrounding compartments, front Stories A peak determining step for determining a data value of one section as a peak value and determining a position of the one section as a peak position;
Selecting a section of the plurality of sections with an arbitrary selection method when a plurality of sections having data values determined as peak values are adjacent or adjacent to each other. Image processing method. The peak determining step determines the data value of the one partition as a maximum peak value when the data value of the one partition is greater than or equal to all the data values of the surrounding partitions, and the one partition 2. The peak detection image processing method according to claim 1, further comprising: determining a data value of the one partition as a minimum peak value when the data value of the first partition is smaller than or equal to all the data values of the surrounding partitions . Claim 1 or 2 peak detection image processing method according to run for all partitioning step and the peak determining step for the comparison operation. The gray image is a two-dimensional rectangular image;
Each section is formed by dividing each of the vertical and horizontal directions of the rectangular image into at least three,
In the comparison calculation step, when the one section is a section that is not in contact with the outer periphery of the rectangular image, the data value of the one section is compared with the data values of the eight sections surrounding the one section. When the one section is a section that is in contact with the outer periphery of only one of the vertical and horizontal directions of the rectangular image, the data value of the one section is set to five pieces surrounding the one section. Comparing with the data value of the section, if the one section is a section at the corner of the rectangular image, the data value of the one section is compared with the data values of the three sections surrounding the one section The peak detection image processing method according to any one of claims 1 to 3 . The peak detection image processing method according to any one of claims 1 to 4, further comprising a step of counting the number of peak values determined by the peak determination step. The grayscale image is optically acquired by an optical image detector such as an imaging tube or a CCD,
The peak detection image processing method according to claim 1, wherein the grayscale data and the position data of the grayscale image data group are generated as digital values from the acquired grayscale image.   The peak detection image processing method according to any one of claims 1 to 6, further comprising a step of recording the determined peak value and peak position.   A peak detection image processing program for causing a computer to execute the peak detection image processing method according to any one of claims 1 to 7. 1 in a grayscale image data group including grayscale data representing gray levels of each section obtained by dividing a grayscale image corresponding to a distribution in a space of an arbitrary physical quantity into a plurality of small sections in association with position data representing the position of each section. Means for individually comparing the grayscale data values of one section with the respective grayscale data values of all surrounding sections surrounding the one section;
Said one if the data value of the partition is greater than or equal to all of the data values of the surrounding compartments, or the one compartment of the data values of all the data values smaller than or equal if the surrounding compartments, front Stories Peak determining means for determining a data value of one section as a peak value and determining a position of the one section as a peak position;
Means for selecting one of the plurality of sections by an arbitrary selection method when a plurality of sections having data values determined as peak values are adjacent to or adjacent to each other. Image processing device. The peak determining means determines the data value of the one partition as a maximum peak value when the data value of the one partition is greater than or equal to all the data values of the surrounding partitions, and the one partition The peak detection image processing apparatus according to claim 9, wherein the data value of the one partition is determined to be a minimum peak value when the data value of is less than or equal to all the data values of the surrounding partitions .